Red Hat Bugzilla – Bug 790298
[glusterfs-3.3.0qa21] - memory leak in glusterfsd process
Last modified: 2016-05-31 21:55:42 EDT
Description of problem:
Memory leak in glusterfsd. valgrind reports 371,699,852 bytes of definitely lost.
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1. Create and start a volume.
2. Attach valgrind to a glusterfsd process.
leak summary from valgrind
==3210== LEAK SUMMARY:
==3210== definitely lost: 371,699,852 bytes in 1,205,625 blocks
==3210== indirectly lost: 68,454 bytes in 2,013 blocks
==3210== possibly lost: 41,852,439 bytes in 4,941 blocks
==3210== still reachable: 36,684 bytes in 70 blocks
==3210== suppressed: 0 bytes in 0 blocks
==3210== Reachable blocks (those to which a pointer was found) are not shown.
==3210== To see them, rerun with: --leak-check=full --show-reachable=yes
==3210== For counts of detected and suppressed errors, rerun with: -v
==3210== ERROR SUMMARY: 238 errors from 238 contexts (suppressed: 30 from 9)
There should be no memory leak.
I have attached the valgrind log.
Created attachment 561771 [details]
Forgot to attach the valgrind log file. Attaching now...
CHANGE: http://review.gluster.com/2746 (rpc : fix the leak with glusterfs v2 auth structure conversion) merged in master by Anand Avati (firstname.lastname@example.org)
CHANGE: http://review.gluster.com/2744 (protocol xdr: remove 'path<>') merged in master by Anand Avati (email@example.com)
CHANGE: http://review.gluster.com/2756 (protocol: remove the 'path<>' from rename() and link()) merged in master by Anand Avati (firstname.lastname@example.org)
With glusterfs-3.3.0qa43, I'm not seeing the memory leak I observed with the qa21 version. Moving to verified.